2

How do you calculate the 99.5 percentile using Math.NET?

98 percentile is

MathNet.Numerics.Statistics.Statistics.Percentile(valuesInDecile, 100 - 2)

99 percentile is

MathNet.Numerics.Statistics.Statistics.Percentile(valuesInDecile, 100 - 1)

For a non integer p-value it says using quartile function. How should this be used as it returns NaN for

MathNet.Numerics.Statistics.Statistics.Quantile(valuesInDecile, 100 - 0.5f)

My data set on which I am trying to run this:

0.0353737
0.0009659
0.0005655
0.0053452
0.0402773
0.0018171
0.0193516
0.0003455
0.0505242
0.0539421
0.0363619
0.0293928
0.0356062
0.0352433
0.0577016
0.0529182
0.0015018
0.0436227
0.0042247
0.0157081
0.0112555
0.0061088
0.0319852
0.020557
0.0106685
0.029806
0.0053733
0.0965764
0.0001344
0.0033052
0.0080388
0.0189088
0.024258
0.0254075
0.0852993
0.0055825
0.0712195
0.0682945
0.013282
StayOnTarget
  • 11,743
  • 10
  • 52
  • 81
sofsntp
  • 1,964
  • 23
  • 34

1 Answers1

2

Try:

MathNet.Numerics.Statistics.Statistics.Quantile(valuesInDecile, 1d - 0.005d)

Instead of grouping into 4 or 100 boxes, quantiles generalize the concept to an infinite number of boxes and thus to arbitrary real numbers tau between 0.0 and 1.0, where 0.0 represents the minimum value, 0.5 the median and 1.0 the maximum value.
Quantiles | Descriptive Statistics

Richard Deeming
  • 29,830
  • 10
  • 79
  • 151
  • That data works fine for me: [demo](https://dotnetfiddle.net/fOYVhU) - the `Percentile` and `Quantile` calls return the same value for 99, 99, and 100 (0.98, 0.99, 1.00); and `Quantile` for 0.95 returns `0.08107536`. – Richard Deeming Oct 04 '21 at 15:11